Hi,all
I've found a strange bug in DFB's inputdrivers(keyboard.c), it seemed that
the DFB driver dispatched more keypress event than real.
That is to say, when I press a key (let's say 'a') down and hold for 1
second to realize the 'repeat' function, then in the linux kernel debug, I got
maybe 9 times 'a' key press interrupt occured, but in the DFB's keyboard
driver, I got more than 7 times 'a' key press event(the amount is 60~90 times
maybe).
And, I've changed the function keyboardEventThread()(in keyboard.c) as
below,
static void*
keyboardEventThread( DirectThread *thread, void *driver_data )
{
... ...
/* Read keyboard data */
while ((readlen = read (data->vt->fd, buf, 64)) >= 0 || errno == EINTR) {
int i;
printf("readlen=%d, evt.key_code=0x%x\n", readlen, buf[0]); //wdl
direct_thread_testcancel( thread );
for (i = 0; i < readlen; i++) {
DFBInputEvent evt;
evt.type = ((buf[i] & 0x80) ?
DIET_KEYRELEASE : DIET_KEYPRESS);
evt.flags = DIEF_KEYCODE;
evt.key_code = buf[i] & 0x7f;
//dfb_input_dispatch( data->device, &evt ); // rem here, so
print is ok!
keyboard_set_lights( data, evt.locks );
}
if (readlen <= 0)
usleep( 200000 ); //wdl, 2000
}
......
}
Then, I got the right times keypress event print info, so I think maybe the
function dfb_input_dispatch() is the bug area, but I can't find how it bugged!
Anybody suggestion? Thanks in advance!
Sean Wang
wang_don...@hotmail.com
_______________________________________________
directfb-users mailing list
directfb-users@directfb.org
http://mail.directfb.org/cgi-bin/mailman/listinfo/directfb-users
